Created bymario-alka.dePowered bygodcore.denoob2claw.detricoma.de
Contributed byMario AlkaNVIDIA

NVIDIA-Nemotron-3-Nano-4B

Performance benchmark · measured on 27.07.2026 08:48

Benchmark-IDrun-20260727-090153-aeee69
Timebench 3 - Kombi (Prefill + Generation)Dense4BRuntime: llama.cppQuantisierung: Q4_K_M
Generation676,42tok/s
Prefill5.366,97tok/s
Time to First Token3.579,00ms
Total duration31,03s
Concurrency5parallel
Ranking in the field
3of 17 systems

Performance benchmark · Primary metric: Generation-Speed (tok/s) · 5× concurrent

This run is better than 88 % of all comparable systems.
Generation 676,4 tok/s
+144 % vs Ø 277,5
Prefill 5.367,0 tok/s
+19 % vs Ø 4.502,5
Time to First Token 3.579 ms
-83 % vs Ø 20.720
Distribution in the field7 – 974 tok/s
Ø 278 Median Dieser Lauf

Wie schlägt sich dieser Benchmark mit anderen Modellen?

How does this benchmark compare on other GPUs?

Same model on different hardware · 5× concurrent · Generation (tok/s)

Hardware

GPU: NVIDIA GeForce RTX 5070 Ti · 16 GB VRAM
CPU: AMD Ryzen Threadripper PRO 5975WX 32-Cores
RAM: 247 GB
Mainboard: ASUSTeK COMPUTER INC. Pro WS WRX80E-SAGE SE WIFI

Setup

Runtime: llama.cpp
Quantization: Q4_K_M
Model: NVIDIA-Nemotron-3-Nano-4B

Configuration

benchmark-konfiguration — run-20260727-090153-aeee69
# LLM-Benchmark Konfiguration # Modell : NVIDIA-Nemotron-3-Nano-4B # Engine : llama.cpp # Run-ID : run-20260727-090153-aeee69 # GPU : NVIDIA GeForce RTX 5070 Ti # CPU : AMD Ryzen Threadripper PRO 5975WX 32-Cores # RAM : 247 GB bench@llm-benchmark:~$ /home/godcore/llama.cpp/build/bin/llama-server \ -m /home/godcore/.cache/huggingface/hub/models--DevQuasar--nvidia.NVIDIA-Nemotron-3-Nano-4B-BF16-GGUF/snapshots/86e3ddc0b34d94852cae725ade1b9b015b516f38/nvidia.NVIDIA-Nemotron-3-Nano-4B-BF16.f16.gguf.Q4_K_M.gguf \ --alias NVIDIA-Nemotron-3-Nano-4B \ --host 0.0.0.0 \ --port 8000 \ -ngl 999 \ -c 16384 \ -np 4 \ --jinja
Engine?Die Inferenz-Software, die das Modell ausliefert (z.B. vLLM oder llama.cpp). Sie bestimmt Geschwindigkeit, unterstuetzte Modellformate und welche Parameter ueberhaupt verfuegbar sind.llamacpp
Modellalias?Der Name, unter dem das Modell ueber die API angesprochen wird. Genau dieser Wert muss im Request-Feld 'model' stehen.NVIDIA-Nemotron-3-Nano-4B
Kontextlaenge?Maximale Anzahl Tokens (Eingabe + erzeugte Ausgabe zusammen), die das Modell pro Anfrage verarbeiten kann.16384
Modellpfad?Pfad zur GGUF-Modelldatei, die geladen und ausgeliefert wird./home/godcore/.cache/huggingface/hub/models--DevQuasar--nvidia.NVIDIA-Nemotron-3-Nano-4B-BF16-GGUF/snapshots/86e3ddc0b34d94852cae725ade1b9b015b516f38/nvidia.NVIDIA-Nemotron-3-Nano-4B-BF16.f16.gguf.Q4_K_M.gguf
GPU-Layer?Anzahl der auf die GPU ausgelagerten Modell-Layer. Hoeher = mehr VRAM und schneller; der Rest laeuft auf der CPU. 999 = alles auf GPU.999
Kontext?Groesse des Kontextfensters in Token. 0 = der beim Training verwendete Kontext des Modells.16384
np4

All benchmarks of this model To leaderboard

Model comparison

NVIDIA-Nemotron-3-Nano-4B on various hardware

All published performance runs of this model – each bubble a variant: position = prefill (X) × generation (Y), bubble size = number of runs. Closer to the top right = faster. ★ Marked gold = this benchmark.

GPUby graphics card

1.5541.1667773890,001.9003.8005.700Prefill (tok/s)Generation (tok/s)AMD Radeon AI PRO R9700 - 261,2 tok/s Generation, 3.495 tok/s Prefill, TTFT 2.882 ms (3 Laufe)AMD Radeon AI PRO R97...AMD Radeon 8060S Graphics - 113,4 tok/s Generation, 1.903 tok/s Prefill, TTFT 5.505 ms (3 Laufe)AMD Radeon 8060S Grap...CPU-only - 8,0 tok/s Generation, 63 tok/s Prefill, TTFT 139.289 ms (3 Laufe)CPU-onlyNVIDIA GeForce RTX 5070 Ti - 1.197,0 tok/s Generation, 4.606 tok/s Prefill, TTFT 4.936 ms (3 Laufe) | DIESER LAUF★ NVIDIA GeForce RTX 50...
★ NVIDIA GeForce RTX 5070 Ti 1.197,0 tok/s this runAMD Radeon AI PRO R9700 261,2 tok/sAMD Radeon 8060S Graphics 113,4 tok/sCPU-only 8,0 tok/s

CPUby processor

1.5541.1667773890,001.9003.8005.700Prefill (tok/s)Generation (tok/s)AMD Ryzen Threadripper PRO 7955WX 16-Cores - 261,2 tok/s Generation, 3.495 tok/s Prefill, TTFT 2.882 ms (3 Laufe)AMD Ryzen Threadrippe...AMD RYZEN AI MAX+ 395 w/ Radeon 8060S - 113,4 tok/s Generation, 1.903 tok/s Prefill, TTFT 5.505 ms (3 Laufe)AMD RYZEN AI MAX+ 395...Intel(R) Xeon(R) CPU E5-4620 v2 @ 2.60GHz - 8,0 tok/s Generation, 63 tok/s Prefill, TTFT 139.289 ms (3 Laufe)Intel(R) Xeon(R) CPU ...AMD Ryzen Threadripper PRO 5975WX 32-Cores - 1.197,0 tok/s Generation, 4.606 tok/s Prefill, TTFT 4.936 ms (3 Laufe) | DIESER LAUF★ AMD Ryzen Threadrippe...
★ AMD Ryzen Threadripper PRO 5975WX 32-Cores 1.197,0 tok/s this runAMD Ryzen Threadripper PRO 7955WX 16-Cores 261,2 tok/sAMD RYZEN AI MAX+ 395 w/ Radeon 8060S 113,4 tok/sIntel(R) Xeon(R) CPU E5-4620 v2 @ 2.60GHz 8,0 tok/s

MBby mainboard

1.5541.1667773890,001.9003.8005.700Prefill (tok/s)Generation (tok/s)ASUSTeK COMPUTER INC. Pro WS WRX90E-SAGE SE - 261,2 tok/s Generation, 3.495 tok/s Prefill, TTFT 2.882 ms (3 Laufe)ASUSTeK COMPUTER INC....Bosgame AXB35-02 (BeyondMax Series) - 113,4 tok/s Generation, 1.903 tok/s Prefill, TTFT 5.505 ms (3 Laufe)Bosgame AXB35-02 (Bey...Dell Inc. PowerEdge R820 - 8,0 tok/s Generation, 63 tok/s Prefill, TTFT 139.289 ms (3 Laufe)Dell Inc. PowerEdge R...ASUSTeK COMPUTER INC. Pro WS WRX80E-SAGE SE WIFI - 1.197,0 tok/s Generation, 4.606 tok/s Prefill, TTFT 4.936 ms (3 Laufe) | DIESER LAUF★ ASUSTeK COMPUTER INC....
★ ASUSTeK COMPUTER INC. Pro WS WRX80E-SAGE SE WIFI 1.197,0 tok/s this runASUSTeK COMPUTER INC. Pro WS WRX90E-SAGE SE 261,2 tok/sBosgame AXB35-02 (BeyondMax Series) 113,4 tok/sDell Inc. PowerEdge R820 8,0 tok/s

ENGby engine

1.3171.2571.1971.1371.0772.3662.4662.5672.668Prefill (tok/s)Generation (tok/s)llama.cpp - 1.197,0 tok/s Generation, 2.517 tok/s Prefill, TTFT 38.153 ms (12 Laufe) | DIESER LAUF★ llama.cpp
★ llama.cpp 1.197,0 tok/s this run

DRVby driver

1.3171.2571.1971.1371.0772.3662.4662.5672.668Prefill (tok/s)Generation (tok/s)unbekannt - 1.197,0 tok/s Generation, 2.517 tok/s Prefill, TTFT 38.153 ms (12 Laufe)unbekannt
unbekannt 1.197,0 tok/s
💰 Economics

Economics of this run

Operating cost, TCO and comparison with the next-best runs of the same model at identical concurrency (5× concurrent). Methodology →

⚙️ ConfigurationAll metrics and charts below follow these settings – based on a 24-month runtime.Save to URLReset
⚡ Electricity price EUR/kWh
⚙️ System utilization 100 %
🖥️ Acquisition EUR
🔌 Idle 10 W
⚡ TDP 10 W
☁️ External LLM (API)
Electricity0.30 EUR/kWh
Avg power (incl. idle)10 W missingBoard 10 W full load
Avg cost / hourEUR 0.0030
Electricity / 1M tokensEUR 0.0012
Token / kWh243.51M
Acquisition (system)EUR 2,096 missingRAM EUR 1,976 · PSU EUR 120
Electricity (2 years)
TCO (2 years)EUR 2,149
Output tokens (2 years)42.66B
☁️ External LLM (API) – comparison
External LLM cost (2 years)
Savings vs. external (2 years)
No power draw measured – values estimated from GPU TDP + CPU (idle + 15 %) + board.

All values above and the charts below take the configured system utilization into account: at X% the system generates only X% of the time, the rest it idles (10 W). Cost per hour drops (more idle), cost per token rises.

Cost over 2 years – electricity only

Cost over 2 years – incl. acquisition (TCO)

Speed vs. tokens per euro

Euro per 1M tokens

Comparison with up to 3 next-best runs of this model at the same concurrency (at least one on different hardware). Power = GPU TDP + CPU (idle + 15 %) + board (estimated), acquisition = full system (GPU + CPU + board + RAM + PSU), prices = stored market prices.

Contributed by

Mario Alka Administrator

@marioalka

Ich bin Unternehmer, Softwareentwickler und KI-Enthusiast. Seit vielen Jahren entwickle ich Unternehmenssoftware und beschäftige mich inzwischen fast täglich mit lokalen LLMs, KI-Agenten und leistungsfähiger KI-Hardware.

Mit LLM-Benchmark.de möchte ich eine Plattform schaffen, auf der Modelle, GPUs und Agenten objektiv und reproduzierbar miteinander verglichen werden.